基于最大熵原理的油气层流体性质识别

IDENTIFICATION FOR HYDROCARBON RESERVOIR FLUID PROPERTY BASED ON THEORY OF MAXIMUM ENTROPY

针对目前各种地层流体性质识别方法的模糊性和不确定性，基于信息论中Jaynes最大信息熵原理提出了一种地层流体性质定量评价模型。利用现场试油资料，将油气层、水层以及干层的各种测井曲线特征值作为地层流体性质判别核心。同时，为避免人为因素干扰，拟选用Shannon熵的方法并利用数据自身信息客观地确定熵权，将待判的各种测井评价参数代入熵最大原理的定量评价模型，计算最小的熵值进而识别流体。该方法既能识别地层流体性质，还可以识别岩性，具有较高的识别精度。

Aiming at ambiguity and uncertainty of current fluid property identification methods for various formations, formation fluid property quantitative evaluation model based on maximum entropy theory is proposed according to Jaynes maximum message entropy theory. Based on field oil testing data, several well logging feature values of hydrocarbon layer, water layer and dry layer are considered as determination core for formation fluid property. Meanwhile, entropy weight is determined objectively by choosing Shannon entropy method by using data in order to avoid artificial interference. Each well logging parameter to be determined is substituted in the quantitative evaluation model of maximum entropy theory to calculate minimum entropy value so as to identify fluid. This method can identify both formation fluid property and lithology with high identification accuracy.